news and views
May the force be with you: Myosin-X
in phagocytosis.
Philippe Chavrier
Engulfment of pathogens by phagocytosis requires the coordination of actin assembly and progressive ‘zippering’ of pseudopodial membranes around the particle. Recent work shows that Myosin-X, which binds
phosphatidylinositol-3-OH kinase (PI(3)K) products through its pleckstrin homology (PH) domains, is required
for phagocytosis, thereby providing a molecular basis for the function of PI(3)K in pseudopod extension.
n higher organisms, phagocytosis is
essential for eliminating infectious agents
and for the scavenging of dead cells1,
whereas in lower unicellular organisms,
phagocytosis is associated with food
uptake. The phagocytic process can be
divided into sequential events, starting with
the recognition of the particle by dedicated
receptors on the phagocytic cell surface.
Among the best-characterized phagocytic
receptors are opsonic receptors, and particularly Fcγ receptors (FcγRs). Clustering of
FcγRs triggers a local and oriented polymerization of actin filaments2 that causes
protrusion of the plasma membrane and
wrapping of the particle within
pseudopods. A contractile force is then
generated to pull the particle into the cell,
and the enclosed particle is finally degraded by lysosomal hydrolases after fusion of
the phagosome with compartments of the
endocytic pathway. Significant advances
have recently been made in the molecular
definition of the components of the
phagocytic signal. Among several signalling molecules that are recruited to and
are required at the site of phagocytosis3,
PI(3)K seems to be essential for pseudopod extension and phagosome closure.
However, in the absence of precisely identified ligand(s) for phosphatidylinositol3,4,5-trisphosphate (PtdInsP3; the product
of PI(3)K) at the phagocytic site, it has been
difficult to assign a definitive function to
PI(3)K in phagocytosis. Work presented in
this issue4 has now identified a recently discovered myosin, Myosin-X, as a downstream effector of PI(3)K in FcγR-signalling,
suggesting new roles for PI(3)K and
PtdInsP3 in phagocytosis.
FcγRs are characterized by the presence
of pairs of tyrosine residues in the receptor
(or its accessory γ chain) cytoplasmic
region. After binding of a particle, receptor
clustering facilitates phosphorylation of
these tyrosine residues by Src family
kinase(s), and brings Syk, a tyrosine kinase
with two tandem src homology 2 (SH2)
domains, to the phagocytic site. In turn, Syk
recruits and activates PI(3)K (ref. 3). By
transfecting a macrophage cell line with
I
VAMP3
Recycling endosomes
Figure 1 A model for PI(3)K-dependent function of Myosin-X during phagocytosis.
Clustering of FcγγRs (not shown for clarity) at the particle attachment site triggers various signalling events that result in filament actin assembly (blue) and accumulation of
PtdInsP3 (red) through the activation of PI(3)K (not shown). Myosin-X (represented as
dimers of heavy chains in purple) is recruited to the forming phagosome through the
interaction of its PH domains (purple dots) with membrane PtdInsP3. The motor head
domain of Myosin-X is engaged on actin filaments and moves towards the barbed ends
of filaments (indicated by arrowheads) facing the tips of the growing pseudopods. By
this dual interaction, Myosin-X would be able to couple actin polymerization and pseudopod extension (black arrows). Membranes required for pseudopod extension are provided by insertion of recycling endosome membranes enriched in the SNARE VAMP3 (light
blue) at the site of phagocytosis.
green fluorescent protein-tagged PH
domains that bind PI(3)K products,
Grinstein and coworkers have recently confirmed that PtdInsP3 rapidly accumulates at
sites of phagocytosis and disappears after
the phagosome has been sealed off from the
plasma membrane5. The evidence that
PI(3)K activity was required for phagocytosis came from experiments using the PI(3)K
inhibitors wortmannin (Wtn) and
LY294002 (refs 6,7). Both drugs significantly reduced the ingestion of large particles
(greater than or equal to 4 µm), whereas
uptake of smaller particles was unaffected7. A
surprising finding of these studies was that,
NATURE CELL BIOLOGY VOL 4 JULY 2002 http://cellbio.nature.com
© 2002 Nature Publishing Group
in contrast with its role in growth factorinduced actin cytoskeleton reorganization
and membrane ruffling, PtdInsP3 production was not required for actin assembly
during FcγR-mediated phagocytosis6,7.
If limitation of pseudopod extension
and inhibition of phagosome closure by
Wtn/LY294002 was not the result of inhibition of cytoskeletal processes, it raises the
question of what it could be caused by.
Based on the observation that Wtn also
abolished the spreading of macrophages on
antibody-coated surfaces (the so-called
‘frustrated phagocytosis’ model), and the
concomitant exocytosis of membrane from
E169
news and views
intracellular pools, Greenberg and colleagues
have suggested that PI(3)K could regulate the
focal delivery of internal membranes to the
plasma membrane, a process thought to contribute to pseudopod extension as a compensatory mechanism for the loss of membrane
taken up into phagosomes7. Accordingly, the
severity of inhibition by Wtn/LY294002
increased with particle size and hence, with
the amount of membrane required to complete engulfment. This conclusion was also
supported by evidence that indicated a
requirement for components of the general
membrane fusion machinery, such as soluble NSF attachment protein receptor
(SNARE) proteins and the ATPase NSF in
phagocytosis8. More recently, other studies9
have examined the exocytosis of a green fluorescent protein (GFP)-tagged fusion of
VAMP3, a SNARE protein that is localized
to recycling early endosomes, which function as intermediates on the receptor recycling pathway. During FcγR-mediated
phagocytosis, VAMP3-positive recycling
endosomes fused with the plasma membrane in a polarized fashion, resulting in
accumulation of GFP–VAMP3 at the
phagocytic site (see Fig. 1). Wtn was found
to inhibit the exocytosis of markers of recycling endosomes during phagocytosis10.
These observations argue that recycling
endosomes may function as a pool of internal membrane that can be readily mobilized
to the cell surface in a PI(3)K-dependent
manner as a fundamental part of the process
of pseudopod extension.
Other studies have suggested the possibility that PI(3)K may be involved in the generation of a contractile activity to complete
phagosome closure6,11. In earlier studies, the
forces produced during engulfment of yeast
particles by phagocytes were measured.
Using a micromechanical method, an alternate phase of engulfment and membrane
extension occurring without contraction
was observed, before a contraction phase
that started abruptly and concomitant to
particle ingestion12. Furthermore, dumbellshaped erythrocytes were observed in another study that examined macrophages
attempting to ingest a single erythrocyte.
The two bulbous ends of the erythrocyte,
each enclosed in a phagosome within the
two adjacent macrophages, were still connected by a thin membrane stalk. In the
presence of Wtn/LY294002 or butanedione
monoxime (BDM, an inhibitor of myosins),
constricted erythrocytes were absent11. This
argues that PI(3)K is involved in the generation of contractility restricted to the
pseudopod margin and in the closure of the
phagosome through a ‘purse-string-like’
mechanism. Several myosin isoforms
(Myosin-Ic, -II, -Ixb and -V) that show a
differential distribution at nascent phagosomes could conceivably control the generation of force during phagocytosis11,13,14.
The missing piece of the puzzle identified
E170
by Cox and colleagues is Myosin-X (ref. 4), a
newly discovered myosin characterized by a
tail domain containing three PH domains15.
Myosin-X is the founding member of a new
class of unconventional myosins that have a
similar motor domain to those of conventional class II myosins, but with unique
structural tail domains that confer classspecific functions16. Human Myosin-X is a
2058-amino-acid protein with head and tail
domains separated by a region predicted to
form a coiled coil, suggesting that MyosinX heavy chains exist as dimers15. The most
unusual feature is the presence of three PH
domains, the second of which interacts
with PI(3)K products17. The carboxy-terminal end contains a MyTH4 (myosin tail
homology 4) and a FERM (4.1, ezrin, radixin, moesin) domain that are conserved in
the tail region of other unconventional
myosins. As expected, Myosin-X binds
F-actin in an ATP-sensitive manner, and is
enriched in actin-rich protrusions, including
lamellipodia and filopodia, in several cell
types15. In a previous issue18, it was reported
that Myosin-X accumulates at the tips of
filopodia, in agreement with the recent
demonstration that Myosin-X is a motor that
moves classically toward the barbed ends of
actin filaments19. Moreover, overexpression
of Myosin-X resulted in a fivefold increase in
the density of filopodia at the cell perimeter18.
These findings strongly argue for a function
of Myosin-X in processes that induce the
protrusion of the plasma membrane.
In this issue, Cox and colleagues show
that macrophage Myosin-X accumulates at
the phagocytic site with similar kinetics to
F-actin, and importantly, this occurs in a
PI(3)K-dependent manner (that is, recruitment is inhibited in the presence of Wtn). A
series of experiments in which truncated
fragments of Myosin-X were expressed in
macrophages allowed the authors to establish the role of Myosin-X in phagocytosis.
Expression of a Myosin-X-tail construct,
comprising the three PH domains, the
MyTH4 and the FERM domain, resulted in
a strong (~75%) inhibition of phagocytosis. However, a similar tail construct with a
point mutation in the second PH domain,
which abolishes PtdInsP3 binding, was not
recruited to the phagocytic site, and accordingly did not inhibit phagocytosis. The
requirement for Myosin-X in the phagocytic process could be further demonstrated by
loading macrophages with antibodies
against the Myosin-X head, resulting in a
significant inhibition of particle ingestion.
Together with the observation that the
Myosin-X-tail did not prevent F-actin
recruitment at the phagocytic site, two
complementary pieces of data suggest that
Myosin-X is required for membrane extension during the phagocytic process. First,
the Myosin-X tail exerted its dominant
inhibitory effect specifically on large particles (6 µm diameter), whereas it was unable
to inhibit ingestion of small particles
(2 µm) that required less membrane.
Second, the tail construct inhibited spreading, but not adhesion, of macrophages on
antibody-coated surfaces.
Myosin-X is a vertebrate-specific
myosin15. However, its tail shares some conserved features with the tail of myosin VII
from Dictyostelium discoideum. Interestingly,
myosin VII-null mutants of Dictyostelium
have a defect in phagocytosis, whereas they
exhibit normal behaviour with respect to
other actin-mediated processes20. Further
evidence suggested that these mutants may
have early adhesion problems, both to the
particle and the substrate21. There are at
least two different mechanisms whereby
Myosin-X could exert its function during
phagocytosis. First, based on compelling
evidence that unconventional class V
myosins function as vesicle transporters in
membrane trafficking, Myosin-X could
function as an actin-based motor to transport membrane cargoes to the forming
pseudopods. Cargoes could include membranes from recycling endosomes that are
delivered to the phagocytic site9. However,
this is rather unlikely because biochemical
evidence suggests that Myosin-X, as
opposed to Myosin V, is not a processive
motor19. Moreover, the authors have shown
that expression of the inhibitory Myosin-Xtail construct does not seem to affect the
recruitment of markers of the recycling
pathway to the phagocytic site. Another
possibility is that Myosin-X could be
involved in generating the forces required
for particle engulfment by pulling on the
actin filament network that is present at the
phagocytic site. Finally, the explanation that
is most favoured by the authors is that by
binding PtdInsP3-enriched membrane and
simultaneously moving along actin filaments, Myosin-X lifts bulk plasma membrane in the direction of the barbed ends
that face the outer margin of the forming
phagosome. Coupled to this process, exocytosis of recycling endosomes provides the
extra membrane required for pseudopod
extension (see Fig. 1).
Over the past five years, components
that contribute to various aspects of the
phagocytic signal have been identified. The
concept that directed actin polymerization
drives the protrusion of the plasma membrane has been extremely fruitful. Along
these lines, proteins such as Rho family
GTPases and their effectors have been
found to participate actively in actin
dynamics during phagocytosis2. However,
another concept has also emerged borrowed from the related problem of cell
motility. This view, initially suggested by M.
Bretscher22, proposed that the polarized
insertion of membrane at the leading edge of
a motile cell pushes it forward. The finding
that Myosin-X functions in phagocytosis by
linking PI(3)K and pseudopod extension,
NATURE CELL BIOLOGY VOL 4 JULY 2002 http://cellbio.nature.com
© 2002 Nature Publishing Group
news and views
and may couple the forward advance of the
actin filament network to the pseudopodial
movement of the plasma membrane, is
important. This result brings together actin
polymerization and membrane delivery to
explain phagocytosis.
Philippe Chavrier is at the Institut Curie – Section
de Recherche, 26 rue d’Ulm, 75005 Paris, France
e-mail: [email protected]
1. Aderem, A. & Underhill, D. M. Annu. Rev. Immunol. 17,
593–623 (1999).
2. Castellano, F., Chavrier, P. & Caron, E. Semin. Immunol. 13,
347–355 (2001).
Cox, D. & Greenberg, S. Semin. Immunol. 13, 339–345 (2001).
Cox, D. et al. Nature Cell Biol. 4, 469–477 (2002).
Marshall, J. G. et al. J. Cell Biol. 153, 1369–1380 (2001).
Araki, N., Johnson, M. T. & Swanson, J. A. J. Cell. Biol. 135,
1249–1260 (1996).
7. Cox, D., Tseng, C. C., Bjekic, G. & Greenberg, S. J. Biol. Chem.
274, 1240–1247 (1999).
8. Hackam, D. J. et al. Proc. Natl Acad. Sci. USA 95, 11691–11696
(1998).
9. Bajno, L. et al. J. Cell Biol. 149, 697–706 (2000).
10. Booth, J. W., Trimble, W. S. & Grinstein, S. Semin. Immunol. 13,
357–364 (2001).
11. Swanson, J. A. et al. J. Cell Sci. 112, 307–316 (1999).
12. Evans, E., Leung, A. & Zhelev, D. J. Cell Biol. 122, 1295–1300 (1993).
3.
4.
5.
6.
13. Allen, L. H. & Aderem, A. J. Exp. Med. 182, 829–840 (1995).
14. Diakonova, M., Bokoch, G. & Swanson, J. A. Mol. Biol. Cell 13,
402–411 (2002).
15. Berg, J. S., Derfler, B. H., Pennisi, C. M., Corey, D. P. & Cheney,
R. E. J. Cell Sci. 113, 3439–3451 (2000).
16. Berg, J. S., Powell, B. C. & Cheney, R. E. Mol. Biol. Cell 12,
780–794 (2001).
17. Isakoff, S. J. et al. EMBO J. 17, 5374–5387 (1998).
18. Berg, J. S. & Cheney, R. E. Nature Cell Biol. 4, 246–250 (2002).
19. Homma, K., Saito, J., Ikebe, R. & Ikebe, M. J. Biol. Chem. 276,
34348–34354 (2001).
20. Titus, M. A. Curr. Biol. 9, 1297–1303 (1999).
21. Tuxworth, R. I. et al. Curr. Biol. 11, 318–329 (2001).
22. Bretscher, M. S. Cell 87, 601–606 (1996).
Leading the way
The ability to chemotax, that is, to sense and move in the direction
of chemical signals, is a feature of a wide variety of eukaryotic cells.
Chemotaxis is important for many biological responses, from the
movement of leukocytes towards sites of infection or inflammation to the aggregation of Dictyostelium discoideum amoebae to
form a multicellular organism. Recent work has firmly established the importance of the phosphatidylinositol 3-OH kinase
(PI(3)K) pathway in mediating directional movement in
response to chemoattractants. Insight into the mechanism that
translates a shallow gradient of chemoattractant into cytoskeletal
polarization and directional movement first came from work
using Dictyostelium cells, and subsequently from studies with
leukocytes and fibroblasts. These studies identified the importance
of localized signalling by demonstrating that green fluorescent
protein (GFP) fusions of a subfamily of pleckstrin homology (PH)
domain-containing proteins, which specifically bind to the phosphoinositide products of PI(3)K, preferentially localized to the
leading edge of chemotaxing cells. These findings strongly suggested that PI(3)K functions at the leading edge of the cell to
mediate directional movement by using its products
PtdIns(3,4,5)P3 and PtdIns(3,4)P2 as second messengers.
Two manuscripts in this issue of Nature Cell Biology (Wang, F.
et al. Nature Cell Biol 4, 513–518 (2002) and Weiner, R. et al.
Nature Cell Biol 4, 509–512 (2002)) report positive feedback loops
in neutrophils that could provide the amplification mechanisms
necessary for the conversion of a shallow extracellular gradient of
chemoattractant into a steep intracellular second messenger gradient. In neutrophils, uniform stimulation with chemoattractant
eventually results in spontaneous polarization. Wang et al.
demonstrate that a membrane-permeable PtdIns(3,4,5)P3 complex can elicit the same response. Using a pharmacological
approach, the authors go on to show that this response is dependent on endogenous PI(3)K activity and requires a Rho family
GTPase activity. Their studies suggest a model for chemotaxis in
which a directional chemoattractant signal results in a small initial activation of PI(3)K, triggering a Rho GTPase-dependent
feedback loop that amplifies the signal, contributing to the
observed steep intracellular PtdIns(3,4,5)P3 gradient.
Weiner et al. provide further evidence that inhibition of
PI(3)K activity impairs the ability to maintain stable pseudopodia,
PtdInsP3 positive feedback loop
PtdInsP3
Rac/
Cdc42
PtdInsP3
PI(3)K
ORION WEINER
resulting in poor chemotactic fidelity. In addition, the authors
provide compelling evidence that actin polymerization at the
leading edge, which drives pseudopod extension and occurs
downstream of PtdIns(3,4,5)P3 accumulation, is in turn necessary for the maintenance of the localized accumulation of
PtdIns(3,4,5)P3 at the leading edge (see figure). Actin dynamics
as part of a positive feedback loop may provide neutrophils with
the ability to spontaneously polarize in response to an initially
diffuse stimulus and start moving, only later homing in on their
target. Thus, the amplification of the response to a chemoattractant gradient by the combination of Rho GTPase and actin feedback loops provides an attractive mechanism for how an initial
small response results in strong cell polarization and persistent
chemoattractant movement. What is unknown in this model are
the mechanisms positioning the initial response that result in the
first accumulation of PtdIns(3,4,5)P3 at the site of the cell closest
to the chemoattractant source. Future studies to define the biochemistry of this step, as well as to flesh out the feedback loops
outlined in these two papers, are necessary to understand how
cells sense and respond to directional signals.
NATURE CELL BIOLOGY VOL 4 JULY 2002 http://cellbio.nature.com
© 2002 Nature Publishing Group
RUEDI MEILI AND RICHARD A. FIRTEL
University of California, San Diego
e-mail: [email protected]
E171